AI and Economic Growth (Deutsche Bank, October 2026): hyperscalers returned 3% in a year against 14% for the S&P 500
Deutsche Bank's Chief Investment Office frames the AI boom through the three drivers of growth: capital, labour and productivity. Its reading is that we are still in the investment and capital-deepening phase. The contribution of computer and software investment to US GDP growth has risen sharply since 2025, while the labour market shows no AI shock yet, with vacancies per unemployed person at about 1.0 against a 2.0 peak in 2022. Estimates for the eventual lift to total factor productivity run from 0.7 percentage points cumulative over a decade (Acemoglu) to 0.8 to 2.8 points a year (IMF).
Markets have already priced the first phase: semiconductor earnings estimates are more than six times their level when ChatGPT launched, yet US hyperscalers returned about 3% over the past year against nearly 14% for the S&P 500, as capex-to-sales ratios climb.
For founders, the bank expects the gains to broaden next into software and services, professional services, health care and financials, which is where the application-layer demand should show up.